Module datarush.analytics
Package com.pervasive.datarush.analytics.pmml

Class PredictiveModelQuality

  • public class PredictiveModelQuality
extends ModelQuality
    A PMML object model for some of the metadata about a predictive (usually regression) model's quality.

    • Field Detail

      • ELEM_THIS

        protected static final String ELEM_THIS
        Constant for the name of the PredictiveModelQuality element.
        See Also:
        Constant Field Values

    • Constructor Detail

      • PredictiveModelQuality

        public PredictiveModelQuality()
        Default constructor for bean-like behavior

    • Method Detail

      • toPMML

        public void toPMML​(Element parent)
        Description copied from class: ModelQuality
        Build this ModelQuality as a PMML element with a given parent
        Specified by:
        toPMML in class ModelQuality
        Parameters:
        parent - the future parent of the element

      • parse

        protected void parse​(Element toParse)
        Description copied from class: ModelQuality
        Initialize this object's state from a PMML element with the appropriate type
        Specified by:
        parse in class ModelQuality
        Parameters:
        toParse - The element with the state

      • getR_squared

        public Double getR_squared()
        Get the r-squared attribute, a measure of the amount of variance in the target variable explained by a model. It ranges from 0.0 (the model explains nothing) to 1.0 (the model is a perfect predictor).
        Returns:
        the "r-squared" attribute for the model, or null if it is not included.

      • getMeanAbsoluteError

        public Double getMeanAbsoluteError()
        Get the meanAbsoluteError, the mean of the absolute values of the predictive errors on that dataset.
        Returns:
        the "meanAbsoluteError" attribute for the model, or null if it is not included.

      • getRootMeanSquaredError

        public Double getRootMeanSquaredError()
        Get the rootMeanSquaredError, the square root of the mean of the squares of the predictive errors on the dataset.
        Returns:
        the "rootMeanSquaredError" attribute for the model, or null if it is not included.

      • getAIC

        public Double getAIC()
        Get the Akaike Information Criterion, a measure of the relative goodness of fit of a statistical model.
        Returns:
        the "AIC" attribute for the model, or null if it is not included.

      • getBIC

        public Double getBIC()
        Get the Bayesian Information Criterion, a measure of the relative goodness of fit of a statistical model which penalizes the number of parameters more strongly than AIC.
        Returns:
        the "BIC" attribute for the model, or null if it is not included.

      • getDegreesOfFreedom

        public Integer getDegreesOfFreedom()
        Get the Degrees of Freedom of the error of the model
        Returns:
        the "degreesOfFreedom" attribute for the model, or null if it is not included.

      • getMeanSquaredError

        public Double getMeanSquaredError()
        Get the Mean Squared Error, the mean of the squares of the predictive errors on that dataset.
        Returns:
        the "meanSquaredeError" attribute for the model-quality element, or null if it is not included.

      • getSumSquaredError

        public Double getSumSquaredError()
        Get the Sum Of Squares (Error) statistic.
        Returns:
        the "sumSquaredError" attribute for the model-quality element, or null if it is not included.

      • getSumSquaredRegression

        public Double getSumSquaredRegression()
        Get the Sum Of Squares (Regression) statistic.
        Returns:
        the "sumSquaredRegression" attribute for the model-quality element, or null if it is not included.

      • getDataUsage

        public PredictiveModelQuality.Usage getDataUsage()
        Get the dataUsage, the relationship between the model and the dataset used to measure its quality. It indicates the phase of model-building during which the model was first exposed to data from that set.
        Returns:
        the "dataUsage" attribute for the model-quality element.

      • getTargetField

        public String getTargetField()
        Get the predicted field on which the quality information was measured.
        Returns:
        the "dataUsage" attribute for the model-quality element.

      • hasExtensionStatistic

        public boolean hasExtensionStatistic​(String name)
        Check whether a certain statistic is available as an extension
        Parameters:
        name - The name of the statistic to check
        Returns:
        true iff the statistic is present and available.

      • getExtensionStatistic

        public double getExtensionStatistic​(String name)
        Retrieve a statistic about the model that is not directly supported by PMML
        Parameters:
        name - the extension statistic to check
        Returns:
        the value of the named extension statistic, or Double.NaN if the extension is not present.

      • getExtensionStatisticNames

        public Set<String> getExtensionStatisticNames()
        Returns:
        a collection of Strings which, when passed to hasExtensionStatistic, return true

      • setTargetField

        public void setTargetField​(String targetField)
        Sets the "targetField" attribute for the model.
        Parameters:
        targetField - the "targetField" attribute for the model.

      • setR_squared

        public void setR_squared​(Double r_squared)
        Sets the "r-squared" attribute for the model. If null, the attribute will be absent.
        Parameters:
        r_squared - the "r-squared" attribute for the model.

      • setMeanAbsoluteError

        public void setMeanAbsoluteError​(Double meanAbsoluteError)
        Sets the "meanAbsoluteError" attribute for the model. If null, the attribute will be absent.
        Parameters:
        meanAbsoluteError - the "meanAbsoluteError" attribute for the model.

      • setMeanSquaredError

        public void setMeanSquaredError​(Double meanSquaredError)
        Sets the "meanSquaredError" attribute for the model. If null, the attribute will be absent.
        Parameters:
        meanSquaredError - the "meanSquaredError" attribute for the model.

      • setRootMeanSquaredError

        public void setRootMeanSquaredError​(Double rootMeanSquaredError)

      • setAIC

        public void setAIC​(Double AIC)

      • setBIC

        public void setBIC​(Double BIC)

      • setDegreesOfFreedom

        public void setDegreesOfFreedom​(Integer degreesOfFreedom)

      • setSumSquaredError

        public void setSumSquaredError​(Double sumSquaredError)

      • setSumSquaredRegression

        public void setSumSquaredRegression​(Double sumSquaredRegression)

      • setDataUsage

        public void setDataUsage​(PredictiveModelQuality.Usage dataUsage)
        Sets the "dataUsage" attribute for the model. If null, the attribute will be set to "training", the default.
        Parameters:
        dataUsage - the "dataUsage" attribute for the model.

      • setExtensionStatistic

        public void setExtensionStatistic​(String name,
                                  double value)
        Sets the value of an attribute not directly supported by PMML. It will appear as an Extension element instead of as an attribute.
        Parameters:
        name - the "name" attribute of the extension
        value - the "value" attribute of the extension